LMZ10504 compensation App Note (SNVA417D)

Thank you; we will look into this

Ah! It looks like some data got dropped upon release of the document. The RCINT and CCINT are internal compensation components. Their values are 105.8kohms and 90pF respectively. The ROUT is simply the output resistance and is calculated as VOUT/IOUT. I hope this helps.

IOUT could be your max current. The LMZ1050x devices are based on the voltage mode architecture. The VOUT doesn't directly influence the compensation. It does have an effect on the calculation of the gain component. What does largely influence is VIN. Hence, in the tables, you will notice that the component values are set for different VIN conditions.

EDIT:

Remember to consider all the capacitors hanging at the output node. We have seen customers connect multiple electrolytic and tantalum capacitors at the output node and close to the load. If these were not considered during the calculation of the compensation components, then you could potentially have an unstable system.

Are you sure that ROUT is defined as VOUT/IOUT? It looks like as definiton for RLoad. As I can understand ROUT is sum of internal Switch's MOSFETs resistance and DC resistance of inductance. And Q of LC loop is depending exactly on this resistance, not on Rload. And wrong ROUT could affect stability. In theory Q higher than 1.2 leads to oscillations.

So is there another information about ROUT for LMZ10503?

What kind of information are you looking for ROUT?
The entire model for the plant transfer function ideally contains MOSFET Rdson, DCR of inductor, ESR of capacitor and the output resistance/impedance. But with all of that in, the model becomes complicated. When you work out the equations (for simplicity without any parasitics but with only include ROUT which is VOUT/IOUT), you will get to the value of Q being equal to ROUT*square-root(COUT/L). The denominator in equation 8 (for Q) has the term ROUT*RESR*COUT. Assuming the output caps are of ceramic make (negligible ESR), the value of that term in comparison to L becomes much smaller and thus negligible. That would bring the equation to the simple form. The addition of ESR was brought about in the equation to account for tantalum caps used in certain applications which skewed the value of ESR. Different write ups will refer to the output resistance with different terms. You will often see RLOAD being used instead of ROUT. In the application note, ROUT refers to output resistance. I have attached a PDF for the calculation of Q for the buck topology.

As an alternative you could download the spice model and run it in TINA to verify any results you want. Or you could use Webench. Is there a reason there are two threads being kept open for the same discussion? I am of the thought that one could be closed and discussion could be continued on just one.

Q_Buck.pdf

Unfortunately I'm also struggling with the LMZ1050x documentation.

The Datasheet link to the app note is wrong, then after finding the app note the formulas are incomplete (Rout definition? Cint, Rcint?)

In addition the webench can't calculate the loop components (some other bug), which makes it difficult to confirm any values

Hello,

I apologize for the delay involved in getting the fixes out. As far as the parameter values go, I have mentioned them in the previous messages in this thread. Maybe I can be of assistance for the other issues you are facing. Please tell me your design requirements.

Regards,
Akshay